Audit Catalog Data Quality for Spare Parts

The audit is not a one-time project. It is a repeatable process that produces a score, a prioritized fix list, and a before-and-after comparison for the next cycle. Run it quarterly, or whenever a major supplier feed is onboarded.

Audit dimensions

The audit process

1. Define the scope

   Decide which product groups to audit. A full-catalog audit is ideal but may not be practical in the first cycle. Start with the product groups that generate the most orders, the most returns, or the most customer complaints about wrong parts.

2. Export the data

   Pull the relevant records: part number, status, description, key attributes, supersession links, applicability records, cross-references. Normalize part numbers and descriptions so comparisons are consistent.

3. Run the supersession checks

   For every part with status “discontinued” or equivalent:

   • Does it have at least one outgoing REPLACED_BY link?
   • Does the chain resolve to a part with status “active”?
   • Does the chain contain any cycles?
   • Are quantity ratios recorded on every link?
   • Is the source recorded on every link?
4. Run the applicability checks

   For every part in the scope:

   • Does it have at least one fitment record?
   • Are the fitment records structured (make, model, serial range) or free text?
   • Are there overlapping serial ranges for the same part and model?
   • Are open-ended ranges flagged for review?
5. Run the description checks

   Sample descriptions across the scope:

   • Do they follow a consistent pattern (type + key attribute + spec)?
   • Are abbreviations consistent (e.g., always “mm” not sometimes “MM” or “millimeter”)?
   • Are there duplicates with different descriptions pointing to the same product?
   • Are descriptions machine-readable or free-text narratives?
6. Run the attribute checks

   For each key specification field:

   • What is the fill rate (% of records with a non-empty value)?
   • Of the filled values, what is the format consistency (same unit, same precision)?
   • Are blanks genuinely “not applicable” for that product type, or are they missing data?
7. Run the cross-reference checks

   For every cross-referenced pair:

   • Do the key specifications match (dimensions, ratings)?
   • Is the relationship type recorded (SAME_AS, EQUIVALENT_TO, SUPPLIED_BY)?
   • Is the source of the cross-reference recorded?
   • Has the pair been verified (auto-matched, human-reviewed, manufacturer-confirmed)?
8. Score and prioritize

   Score each dimension as a percentage. Prioritize fixes by business impact:

   1. Order accuracy — supersession gaps, wrong cross-references (highest priority)
   2. Findability — missing applicability, poor descriptions (medium priority)
   3. Data maturity — attribute fill rates, format consistency (ongoing improvement)

Scoring template

The targets are starting points. Adjust them based on your industry, catalog size and the cost of wrong-part orders. A heavy-equipment distributor where a wrong part means a machine down for a week should aim higher than a general MRO supplier.

Common findings

• Large clusters of discontinued parts with no successor link — often from a supplier who sends a new price list but not a supersession feed.
• Supersession chains that end at another discontinued part — the intermediate successor was itself superseded and nobody extended the chain.
• Parts with “fits all models” applicability instead of structured records — makes will-fit searches too broad to be useful.
• Descriptions that vary between “Bearing 6205-2RS” and “DEEP GROOVE BALL BRG 6205 2RS SEALED” for the same product — breaks deduplication.
• Cross-references loaded from a third-party database without specification verification — some pairs match on number but not on dimensions.

Where Claro fits

Claro runs these checks continuously against the product identity layer. Supersession gaps, broken chains, unverified cross-references and missing applicability records surface as exception queues for the catalog team. The scoring is computed automatically and tracked over time, so each audit cycle shows measurable improvement.

Related

FAQ